ocrs-cli
CLI tool for extracting text from images using the ocrs OCR engine.
Installation
These steps assume you have Rust and cargo installed.
Usage
Extract text from an image and print it to standard output:
The first time that you run the tool it will download default models for text
detection and recognition. You can override this by using the --detect-model <path> and --rec-model <path> flags to set the path to the models.
Layout export
By default ocrs outputs just the extracted text. Specifying the --export-boxes path.json flag will cause text layout information to be produced in JSON
format.
Debug output
If the --debug flag is enabled, ocrs will create an annotated image in the
current directory showing where words and lines were detected, as well as
logging performance information.
Language support
The current version of this tool was trained primarily on English text. Support for more languages is planned for the future.